Research On Preemptive Job-Shop Problem In Steel And Iron Industry And Its Software Solution

According to the job plan, aiming at enterprise's business and when all comes to all, production scheduling combines real production process and recent production capacity to optimize the production plan and to control materials and enery to coordinate the delivery among the machines so that it can make the production balanced and concerted on all the machines and production phases effectively connected and tightly linked up. Therefore under the optimized instruction, the production process would be banlanced, stable and safe in a long term. With the deepening of informationization in the iron and steel industry, these enterprises really need the models and algorithms in line with their production process to support for production decisions.Firstly this paper briefly reviewed the domestic and foreign research status quo in light of production scheduling theory, Job-Shop scheduling problems and the scheduling problems in iron and steel industry. And secondly with regard to the real problem in the production of steel pipes of a large domestic iron and steel enterprise, under the consideration of continuous production, inventory restriction and partial manual scheduling, the problem is abstracted into a Preemptive Job-Shop Scheduling Problem by time slicing. And then the problem's models and algorithms were researched seperatively in mathematical programming and heuristic method.According to the production constraints (storage, partial manual scheduling, consecutive producing) and by aiming at the minimum preemptive times, a mixed-integer programming model was established and could be solved by utilizing ILOG CPLEX software package. However, because of the partial manual scheduling, sometimes the feasible solution cannot be found. Then the heuristic method was put forward after changing the model's aim and constraints. As for the heuristic algorithm, some rules were summed up and a heuristic algorithm was developed. The solution was receptable but it still needed improving. Then in the improved two-phase heuristic algorithm, four rules (based on job sequence, SPT, ERT and EDD) were applied to schedule the slackest production line seperatively for the first phase and the second one was similar to the heuristic algorithm. The solutions were more ideal. What's more, applied to the same case example, the six solutions gained by three algorithms were compared under rough evaluation and the conclusion was that the best one was improved heuristic algorithm (SPT); the better ones were mathematical programming (ILOG Cplex) and improved heuristic algorithm (EDD) as well; and then improved heuristic algorithm (ERT); the worst were improved heuristic algorithm (Job sequence) and heuristic algorithm.Finally, the conclusion is made and the problems requiring further studies are briefly discussed.